view scripts/stats.xmlrpc.py @ 6610:db3f0ba75b4a

Change checkpoint_data and restore_connection_on_error to subtransaction checkpoint_data and restore_connection_on_error used to commit() and rollback() the db connection. This causes additional I/O and load. Changed them to use 'SAVEPOINT name' and 'ROLLBACK TO name' to get a faster method for handling errors within a tranaction. One thing to note is that postgresql (unlike SQL std) doesn't overwrite an older savepoint with he same name. It keeps all savepoints but only rolls back to the newest one with a given name. This could be a resource issue. I left a commented out release statement in case somebody runs into an issue due to too many savepoints. I expect it to slow down the import but....
author John Rouillard <rouilj@ieee.org>
date Sat, 29 Jan 2022 11:29:36 -0500
parents 75da037d1c54
children
line wrap: on
line source

"""Count how many issues use each bpo field and print a report."""
""" sample output: https://github.com/psf/gh-migration/issues/5#issuecomment-935697646"""

import xmlrpc.client

from collections import defaultdict

class SpecialTransport(xmlrpc.client.SafeTransport):
    def send_content(self, connection, request_body):
        connection.putheader("Referer", "https://bugs.python.org/")
        connection.putheader("Origin", "https://bugs.python.org")
        connection.putheader("X-Requested-With", "XMLHttpRequest")
        xmlrpc.client.SafeTransport.send_content(self, connection, request_body)

# connect to bpo
roundup = xmlrpc.client.ServerProxy('https://bugs.python.org/xmlrpc',
                                    transport=SpecialTransport(),
                                    allow_none=True)

# map bpo classes -> propname
# the class is the name of the class (e.g. issue_type, keyword --
# also used in e.g. in https://bugs.python.org/keyword)
# the propname is the name used as attribute on the issue class
# (e.g. issue.type, issue.keywords)
classes = {
    # 'status': 'status',  # skip this
    'issue_type': 'type',
    'stage': 'stage',
    'component': 'components',
    'version': 'versions',
    'resolution': 'resolution',
    'priority': 'priority',
    'keyword': 'keywords',
}

# find the id for the 'open' status
open_id = roundup.lookup('status', 'open')

print(f'* Counting total issues...')
total_issues_num = len(roundup.filter('issue', None, {}))

print(f'* Counting open issues...')
# use this list to filter only the open issues
open_issues = roundup.filter('issue', None, {'status': open_id})
open_issues_num = len(open_issues)

# save the totals in a dict with this structure:
#   totals[propname][open/all][num/perc][name]
# where propname is e.g. 'keyword' and name is e.g. 'easy'
totals = defaultdict(lambda: {'all': {'perc': {}, 'num': {}},
                              'open': {'perc': {}, 'num': {}}})
for cls, propname in classes.items():
    print(f'* Counting <{cls}>...')
    # get the list of ids/names for the given class (e.g. 'easy' is 6)
    ids = roundup.list(cls, 'id')
    names = roundup.list(cls, 'name')
    for id, name in zip(ids, names):
        # filter and count on *all* issues with the given propname
        tot_all = len(roundup.filter('issue', None, {propname: id}))
        totals[propname]['all']['num'][name] = tot_all
        totals[propname]['all']['perc'][name] = tot_all / total_issues_num
        # filter and count on *open* issues with the given propname
        tot_open = len(roundup.filter('issue', open_issues, {propname: id}))
        totals[propname]['open']['num'][name] = tot_open
        totals[propname]['open']['perc'][name] = tot_open / open_issues_num


print(f'Issues (open/all): {open_issues_num}/{total_issues_num}')

# print a list of markdown tables for each bpo class name
for propname in classes.values():
    print(f'### {propname}')
    print('| bpo field | open | all |')
    print('| :--- | ---: | ---: |')
    # pick the dict for the given propname (e.g. keywords)
    proptots = totals[propname]
    names = proptots['open']['num']
    # sort the names (e.g. 'easy') in reverse order
    # based on the number of open issues
    for name in sorted(names, key=names.get, reverse=True):
        # get and print num/perc for all/open issues
        issues_all = proptots['all']['num'][name]
        issues_open = proptots['open']['num'][name]
        perc_all = proptots['all']['perc'][name]
        perc_open = proptots['open']['perc'][name]
        print(f'| {name:20} | {issues_open:>5} ({perc_open:5.1%}) |'
              f' {issues_all:>5} ({perc_all:5.1%}) |')
    # calc and print num/perc for all/open issues
    tot_issues_all = sum(proptots['all']['num'].values())
    tot_issues_open = sum(proptots['open']['num'].values())
    tot_perc_all = sum(proptots['all']['perc'].values())
    tot_perc_open = sum(proptots['open']['perc'].values())
    print(f'| **Total**            | {tot_issues_open:>5} ({tot_perc_open:5.1%}) |'
            f' {tot_issues_all:>5} ({tot_perc_all:5.1%}) |')


Roundup Issue Tracker: http://roundup-tracker.org/